Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

Biogeosciences Discuss., 10, 16137–16171, 2013www.biogeosciences-discuss.net/10/16137/2013/doi:10.5194/bgd-10-16137-2013© Author(s) 2013. CC Attribution 3.0 License.

Open A

ccess

BiogeosciencesDiscussions

This discussion paper is/has been under review for the journal Biogeosciences (BG).Please refer to the corresponding final paper in BG if available.

Extreme dissolved organic nitrogenfluxes in the human impacted PambaRiver, Kerala, IndiaS. Elizabeth David and T. C. Jennerjahn

Leibniz Center for Tropical Marine Ecology, Fahrenheitstraße 6, 28356 Bremen, Germany

Received: 2 August 2013 – Accepted: 1 October 2013 – Published: 18 October 2013

Correspondence to: S. Elizabeth David (shilly.elizabethdavid@zmt-bremen.de) andT. C. Jennerjahn (tim.jennerjahn@zmt-bremen.de)

Published by Copernicus Publications on behalf of the European Geosciences Union.

16137

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

Abstract

Cultural eutrophication of coastal aquatic systems is a major undesired phenomenonof today, which is mainly ascribed to the application of inorganic fertilizers in agricul-ture. Consequently, dissolved inorganic nitrogen (DIN) is considered the major problemand widely studied. However, human interventions also strongly influence the riverine5

dissolved organic nitrogen (DON) concentrations and fluxes. Studies of nutrient inputsfrom tropical river catchments are biased towards DIN, even though they account foronly a portion of the total dissolved nitrogen (TDN) pool, whereas the rest is comprisedof DON and has been largely ignored. The tropical Pamba River was studied becauseof its manifold human activities in the catchment and was sampled during the south10

west monsoon (SWM), north east monsoon (NEM) and the pre monsoon (PM) monthsduring 2010 to 2013. The largest pilgrim center on earth, the Sabarimala temple, lo-cated in the upstream forest is a unique feature of the catchment. Fertilizer application,livestock farming and inadequate sewage treatment are the prevailing land use prac-tices. The goals of this study were to (i) define cause-effect relationships by assessing15

the effect of various human interventions such as the pilgrims, agriculture and sewagedisposal in combination with the seasonal variations in hydrology on the DON concen-trations and fluxes and to (ii) quantify the inputs from respective land use segments.

The global maximum DON concentration (29 302 µM) was measured for the PambaRiver. Pilgrim activities, high population density, agricultural and livestock farming as20

well as the lack of infrastructure for sanitation facilities were the cause for extremelyhigh DON concentrations and fluxes in the plantation and settlement with mixedtree crop (SMT) segments. A DON yield of 745 kg ha−1 yr−1 was calculated for thePamba catchment. The total DON inputs from all quantifiable sources amounted to514 kg ha−1 yr−1 comprising of 69 % of the total Pamba DON yield. In the Pamba River,25

sewage is the major source of DON and the unique Sabarimala pilgrim event accountsfor most of it. Nevertheless, sewage input from the rest of the densely-populated catch-ment is high, which is a common feature of developing countries that lack adequate

16138

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

sanitation and water technology, i.e. in South and Southeast Asia and tropical Africa.Our study shows that DON makes up a significant portion of anthropogenic nitrogenin rivers, in particular in those regions, which are, however, scarce in respective data.It underscores the need for more quantitative studies from densely-populated tropicalriver catchments in order to improve global nitrogen budgets and the assessment of5

the consequences of anthropogenic nitrogen inputs into coastal aquatic systems.

1 Introduction

Rivers play a major role in the transport of nutrients from the terrestrial source to thereceiving water bodies and are affected by both natural and anthropogenic factors.Anthropogenic factors such as high population density, deforestation, construction of10

dams, fertilizer usage, sewage disposal strongly alter the nature and quantity of nu-trient fluxes in the river catchment. Nitrogen (N) is one of the important nutrients forthe ecosystem functioning. Human activities have greatly accelerated N fluxes in theterrestrial biosphere (Jordan and Weller, 1996) and fluxes of N cycled through aquaticsystems (Vitousek et al., 1997). N inputs to the aquatic system have increased sub-15

stantially over the last century because of human activities and, the anthropogenicfixation of reactive nitrogen (190 TgNyr−1) exceeds that of natural biological fixation(110 TgNyr−1, Galloway et al., 2004). The enhanced supply of N lead to various en-vironmental problems such as eutrophication, formation of hypoxic zones and harmfulalgal bloom (Rabalais, 2002; Turner et al., 2003).20

Bouwman et al. (2005) estimated that on a global scale half of the riverine N fluxis derived from anthropogenic activities. In river catchments with high population den-sity and manifold human activities dissolved organic nitrogen (DON) can be the majorN contributing to the total dissolved nitrogen (TDN) pool. Waste water discharge andrunoff from intensive agricultural activity within watersheds are mainly responsible for25

the elevated DON concentrations in drainage waters (Westerhoff and Mash, 2002).Studies have shown that the source and composition of DON influence microbial re-

16139

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

sponse (Seitzinger et al., 2002) and potentially contribute to coastal eutrophication andhypoxia (Glibert et al., 2006; Seitzinger and Sanders, 1997). The dissolved organicforms of nitrogen are largely ignored even though they account for a substantial pro-portion of the total dissolved nitrogen (TDN) in the rivers draining both undisturbed anddisturbed catchments (Perakis and Hedin, 2002; Lovett et al., 2000; Seitzinger et al.,5

2005).Population density is one of the most commonly used parameters for expressing

demographic-human impact on N concentrations and fluxes in rivers (Martinelli et al.,2010). N transport is greater for rivers draining more densely populated basins (Coleet al., 1993). High N export rates are reported from human altered tropical watersheds10

(Boyer et al., 2006; Filoso et al., 2006). Even though much of the world’s populationlive in the developing countries of the tropics, the effects of human activities on river-ine N export are poorly understood (Martinelli et al., 2010), mainly because of a lackof data. South and Southeast Asia are among the regions with the strongest humanmodifications of the coastal zone due to their high population density.15

Large areas in central Asia including India struggle with establishing basic waterservices like clean drinking and sanitation (WHO/UNICEF, 2010). About 80 % of theworld’s population is exposed to high levels of threat to water security. Regions ofintensive agriculture and dense settlement show high water security and biodiversitythreat of which, the Indian subcontinent is one among them (Vörösmarty et al., 2010)20

Rapid increase in the population and the need to meet the increasing demand forirrigation, human and industrial consumption, water resources in many parts of Indiaare getting depleted (Bhardwaj, 2005).

Very little is known about the DON concentrations and fluxes from densely populatedtropical Indian river catchments. In this study the source and transport of DON concen-25

trations and fluxes is determined in the densely populated and human impacted PambaRiver catchment. The unique feature of the Pamba River is the presence of Sabarimalatemple, the largest pilgrim center in the world in its upstream catchment. Besides thelarge pilgrim activity, agricultural land use practices and livestock farming and sewage

16140

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

effluents are the major anthropogenic activities in the Pamba catchment. The mainobjective of the study is to define direct cause-effect relationships by understandingthe effect of various human interventions such as the pilgrims, agriculture and sewagedisposal in combination with the seasonal variations in hydrology on the dissolved or-ganic nitrogen concentrations and fluxes in the Pamba catchment. A segment wise5

DON load and yield was calculated to assess the impacts caused by the individualhuman interventions into the river.

2 Materials and methods

2.1 Study area

The study area is located in the south west coast of India in the state of Kerala. The10

Pamba is one of the 41 west flowing rivers in Kerala originating from the Western Ghatmountains in the east. The Pamba River has a length of 176 km and a total basin areaof 2235 km2. The river originates from Pulachimala in the Western Ghats at an altitudeof 1650 m and flows westwards and debouches in to the Vembanad estuary with anannual discharge of 3.9 km3.15

The Pamba catchment has a population density of 916 individualskm−2 making upa total of 900 000 inhabitants (Census of India, 2011). The climate is dominated by twomonsoons i.e., south west monsoon (SWM) and north east monsoon (NEM) followedby a dry pre monsoon (PM) season. 60 % of the total precipitation is obtained duringthe SWM from mid May/beginning of June to August and 30 % during the NEM from20

October to December and 10 % during the PM from January to mid May (Fig. 1).The Pamba catchment can be categorised into three topographically distinct regions

that are representative of different land use practices i.e., the upstream forest region,midstream settlement with mixed tree crops (SMT) region and downstream paddy cul-tivable region. Major land use categories in the Pamba basin include forest, comprising25

of 44 % of the total land area. Two reservoirs – Pamba and Kakki – with a storage

16141

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

capacity of 493.5×106 m3 are located in the dense forest region and cover 0.9 % ofthe total land area. Sabarimala temple, the largest pilgrimage centre in the world issituated in the upstream catchment. During the pilgrim season i.e., from Novemberto January, about 50 to 60 million pilgrims visit the temple every year (BusinessLine,2009). Forest plantations such as tea, teak, rubber, cardamom and eucalyptus occupy5

14 % of the total land area. Settlement area together with mixed crops such as veg-etables, banana, tapioca, coconut, mango, jack and cashew trees comprises the mid-stream reaches. Settlement with mixed tree crop (SMT) area covers 14 % of the totalland area. Paddy cultivation is the main land use practice in the downstream reachestogether with a small SMT area. Compared to the upstream reaches, the mid- and10

downstream reaches have a higher population density (Table 1, Fig. 2).The Pamba catchment was divided into 8 segments for the segment wise load and

yield estimation. Segments were categorized based on land use and on hydrologicunits (Fig. 2). The upstream forest area is subdivided into 3 regions i.e., dense for-est, temple and plantation regions, respectively. Segment I comprises of minimally dis-15

turbed dense forest region. The temple region is represented by segment II. Segment IIIrepresents mainly the forest area with mixed plantation. Segment IV comprises mainlythe forest and rubber plantation along with a small settlement area. Segment V is themain confluence point of the above segments and represents mainly rubber plantationwith a small SMT region. The Malakkara gauging station divides segments VI and VII.20

Segment VI represents SMT along with a small area of rubber plantation, while SMTand paddy cultivation is the main land use in segment VII. Land use of segment VIIIis similar to that of segment VII and was categorized as a different segment becauseof the merging of the adjacent Achankovil and Manimala rivers with the Pamba Riverforming an inland deltaic region. Individual river contributions were difficult to distin-25

guish in segment VIII.

16142

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

2.2 Sampling and analysis

16 locations were sampled along the river course from the upstream forest to the estu-arine mouth. Monthly sampling was done for 9 months during 2010 to 2012 to cover theSWM, NEM and PM seasons. The sampling locations were selected based on land useand watershed characteristics. The samples were grouped based on the land use cat-5

egories as undisturbed forest (location 1), temple (location 2 to 4), plantation (location5 to 8), SMT (location 9 to 11) and paddy (location 12 to 16) areas, respectively.

For the total dissolved nitrogen (TDN) analysis 10 mL of water were filtered throughsingle use Sartorius syringe filters into a glass vial and fixed with 13 µL of 20 %phosphoric acid. Samples were stored frozen until analysis. Total dissolved nitrogen10

(TDN) was determined in a Shimadzu TOC-VCPH Total Organic Carbon Analyzer witha TNM-1 Total Nitrogen Measuring Unit combusting at 720 ◦C. The detection limit was0.32 mgL−1. Random samples ranging from low to high TDN concentrations were an-alyzed again in the Institute of Biogeochemistry and Marine chemistry, University ofHamburg using TOC-V CSH/CSN Total Organic Carbon Analyzer. The detection limit15

was 0.29 mgL−1. The coefficient of variation of the analysis was < 1.5 %. Dissolved in-organic nitrogen (DIN) were analyzed using a continuous flow analyzing system (SkalarSAN++System). Nitrate+nitrite (NO−

x ) and nitrite (NO−2 ) were determined spectropho-

tometrically and ammonium (NH+4 ) fluorometrically (Grasshoff, 1999). DIN was calcu-

lated by summing up NO−3 , NO−

2 and NH+4 . The relative standard deviation of each20

method was < 3.3 %. Determination limits were 0.075 µM, 0.035 µM and 0.051 µM forNOx, NO2 and NH4, respectively. Dissolved organic nitrogen (DON) was calculated asTDN−DIN.

2.3 Calculation of segment wise river DON budgets

For the budget calculation, the exit location of each segment (end location) was se-25

lected. Discharge data for the Malakkara gauging station were obtained from the Cen-tral Water Commission (CWC), Government of India. Annual flow at the gauging station

16143

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

was calculated based on the monthly discharge data. Discharge per cubic kilometerand DON load and yield of the river were calculated using the following equations.

Discharge in km3 km−2 (Q) = QM/AM (1)

where QM = discharge at the Malakkara gauging station; AM = catchment area tillMalakkara gauging station5

Discharge per segment (Qs) = As ×Q (2)

where As = segment area (km2)

DON load (tyr−1) = C ×Qs/1000 (3)

where C = average concentration of the end location (mgL−1)

DON yield (kgha−1 yr−1) = DON load/As ×100 (4)10

Catchment area is the most important discharge determining factor as has been ob-served in a case study from Kerala. There, annual water discharge correlated signifi-cantly with catchment area (r = 0.78; Chattopadhyay and Chattopadhyay, 2009). Fol-lowing this approach we calculated a discharge (Q) of 0.2 km3 km−2 of catchment areafor the Pamba River, which we then used for segment wise nutrient budget calculations.15

In the absence of robust discharge data from specific land use segments, this methodis perhaps the best approximation.

2.4 Statistics

Data were statistically analyzed using the SigmaPlot 12.3 software. A one-way anaysisof variance (ANOVA) was performed to test the seasonal and spatial variations. Lin-20

ear regression analysis was done on log-transformed data to evaluate the correlationsbetween population density, sewage discharge and DON yield.

16144

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

3 Results

DON concentrations ranged from 36 to 29 302 µM with an average concentration of5940 µM in the river during the whole study period. Both spatial and temporal varia-tions in DON concentrations were observed (Fig. 3). Location 1 in the upstream forestregion exhibited comparatively low values with an average concentration of 371 µM.5

Concentrations were found to be high throughout the catchment, particularly in thetemple, plantation and SMT segments. The maximum concentration of 29 302 µM wasobserved at location 2 during the pilgrim months. In the temple locations DON con-centrations varied widely before, during and after the pilgrim event. Concentration was429 µM in October i.e., before the pilgrim event and increased up to 20 201 µM during10

the event in November and decreased to 64 µM after the event in April in location 3.When compared to other segments, values were high in the plantation and SMT seg-ments during the NEM. The second highest concentration of 28 724 µM was observedin location 6 in the plantation segment. Concentrations were high in the agriculturaldominated downstream segments, particularly in the paddy cultivated segments dur-15

ing the PM, except for high values in the temple locations during January and February.A maximum concentration of 21 014 µM was observed in location 14 during March. Inlocation 13 during April a maximum value of 25 862 µM was displayed and was highcompared to other locations. A maximum value of 11 648 µM was observed in location10 during the SWM. Significant seasonal variations in concentration were observed20

between NEM and PM months (p < 0.050).The minimum annual DON load of the Pamba River of 2244 t was observed in seg-

ment I. A maximum load of 84 929 t was contributed from segment IV, followed by66 982 t from segment III and 43 398 t from segment II, respectively. A minimal loadingof 13 421 t was observed at Segments VII, compared to the other agricultural segments.25

Similar to the load, a minimum DON yield of 90 kgha−1 yr−1was observed at segment I.The highest yield of 1976 kgha−1 yr−1was observed at segment II, followed by yields of1863 kgha−1 yr−1and 1818 kgha−1 yr−1 in segments III and V, respectively. In the down-

16145

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

stream segments VI and VIII, DON yield exhibited a decreasing trend towards the rivermouth (Table 2).

A full segment wise quantification of DON inputs per land use is not possible with-out being able to quantify losses during transport. Because of this missing loss terma simple subtraction of the DON load of an upstream segment from that of the next5

downstream segment can in some instances result in a negative load. Therefore, ab-solute quantification was possible only for segments I, II and IV. Nutrient quantificationof segment II, i.e. from pilgrims was calculated because of very low nutrient input fromthe undisturbed segment I. Segment IV was not influenced by the upstream segmentsI and II and high nutrient inputs therein were from the fertilizer inputs and settlements.10

In these cases, the respective numbers given for the downstream segments can beconsidered semi-quantitative only while, absolute numbers were given for the undis-turbed, temple and plantation segments. Nevertheless, segment wise load and yieldcalculations are the best estimates possible and provide important insights into DONbudgets in human-impacted Pamba river catchment.15

A socio-economic survey was conducted among farmers from the paddy, SMT andrubber plantation regions to collect information about the quality, quantity and time offertilizer application in the agricultural fields. Factamphose, urea and potash were thecommon synthetic NPK fertilizers applied in the fields. The NPK fertilizer loads in theland use segments were calculated by multiplying the rate of fertilizer application with20

the respective land area. Detail survey results are described in (Elizabeth David et al.,2013). Annual synthetic N fertilizer input in the paddy, SMT and rubber plantation re-gions were 3313 t, 4885 t and 1644 t, respectively. Of the total N input, 45 % was com-posed of urea and amounting to 44.3 kgha−1 yr−1. Agricultural land occupies 37 % ofthe total land area in the Pamba catchment. When normalized to the whole catchment25

area, the total synthetic N fertilizer input was 44 kgha−1 yr−1, of which 20 kgha−1 yr−1

was urea.

16146

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

4 Discussion

Both the concentrations and loads of nitrogen in the Pamba catchment are dominatedby dissolved organic nitrogen (99 %). The reasons for the high DON values are dis-cussed below.

4.1 Impacts from the undisturbed land use area5

DON concentrations and yield in the undisturbed forested segment I was low comparedto the values from the human impacted segments (Fig. 3, Table 2). The yield in segmentI in the Pamba catchment was much higher than in other forested catchments in theworld.

DON in the stream water of forested regions can originate from throughfall, leaching,10

decomposition of litter and soil organic matter, plant exudates, and the waste prod-ucts of macro- and microorganisms (Campbell et al., 2000). A study conducted in thewet tropical La Selva forest in Costa Rica quantified the DON flux in the through-fall (9 kgha−1 yr−1), litter leachate (13 kgha−1 yr−1) and in the soil solution (0.4 to3 kgha−1 yr−1) (Schwendenmann and Veldkamp, 2005). Another study conducted in15

three subtropical forests in South China under high air pollution quantified the DONflux in precipitation (17.8 kgha−1 yr−1), throughfall (14.6 to 20.1 kgha−1 yr−1) and soilsolution (6.5 to 16.9 kgha−1 yr−1), respectively. DON accounts for 28 to 40 % of the to-tal dissolved N (Fang et al., 2008). The climate is warm and humid with a mean annualprecipitation and humidity of 1927 mm and 80 %, respectively.20

In the Pamba catchment, the climate is warm and humid. The annual average pre-cipitation rate is higher (2643 mm, Central Water Commission) than reported for thesubtropical forests in South China. Dense forest cover comprising of organic rich for-est loam soil and heavy precipitation rates during monsoon months are characteristicfeatures of segment I. Taking into account the climate conditions, segment I in the25

Pamba catchment is similar to that of the subtropical South China forest. Natural bio-logical nitrogen fixation (BNF) can be an important DON source in the forest segment,

16147

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

as for instance, in the Latin America, Africa and Oceania regions, of the total N 85 %was comprised by BNF (Galloway et al., 2004; Boyer et al., 2006). Due to a lack offield data, the DON from the abovementioned specific sources cannot be quantified.Missing data from other regions with a comparable environmental setting even did notallow an estimate for BNF. It is concluded that the major DON in this forest segment is5

from throughfall, leaching and decomposition of litter and soil organic matter and plantexudates as observed for the South China forests.

4.2 Effect of pilgrim activity

Dissolved organic nitrogen concentrations were extremely high in the temple locations(Fig. 3) during the pilgrim season. Before visiting the Sabarimala shrine pilgrims con-10

ducted ritualistic bathing in the Pamba River. Various activities such as usage of soap,discarding clothes and plastics, urinating and, in many instances defecating on theriver banks were practiced. Similar activities were also reported in different parts of theGanges River in North India, where millions of people gather for the famous Kumbh fes-tival. In 2010, this event was taken place in the Haridwar city in Uttarakhand State and15

41.6 million pilgrims took ritualistic bathing in the river thereby, leading to water qual-ity deteriorations (Tyagi et al., 2012). Biological oxygen demand (BOD) and total sus-pended solid (TSS) concentrations varied before, during and after the event. Before theevent, the mean BOD and TSS concentrations were 2±0.4 mgL−1 and 17±3.7 mgL−1.During the event the values increased up to 6±2.9 mgL−1 and 30±9.7 mgL−1and, after20

the event the values decreased considerably to 2±0.4 mgL−1 and 19±1.9 mgL−1, re-spectively. High numbers of fecal coliforms were observed during the event. Dissolvedoxygen (DO) and pH varied slightly before, during and after the bathing event in theGanges River. In addition to the bathing event, the offering of milk, curd, sugar, flowers,ashes of departed ones, body hairs etc into the river resulted in water quality deterio-25

rations during the event (Tyagi et al., 2012).In the Pamba temple locations, DO and pH values varied considerably before, dur-

ing and after the pilgrim event. Before the event the average DO and pH values were16148

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

5.8±1 mgL−1and 7.3±1. Values decreased to 3.5±1 mgL−1 and 7.1±0 during, and in-creased to 4.2±1 mgL−1and 7.1±1 after the event, respectively. Similarly, the dissolvedinorganic nitrogen (DIN) values also displayed variations before, during and after theSabarimala pilgrim event (Elizabeth David et al., 2013). The same was observed forthe DON where, the value was two to three orders of magnitude higher (20 201 µM)5

during the event than before (429 µM) and after (64 µM) it (Fig. 3).Organic N content in faeces and urine of an adult person is ∼ 1.5 and 1.7 g per day,

respectively, and N in the form of ammonia and urea in urine of an adult person is10.5 g per day (Painter and Viney, 1959). Urea is the main organic N compound of itamounting to 85 % (GeigyScientificTables, 1981). Approximately, 555 000 pilgrims visit10

the temple per day during the pilgrim season. Pilgrim activities are mainly concentredin about 10 km2 area downstream of the Sabarimala shrine. DON flux during the pilgrimseason was calculated by multiplying the individual organic N load in human wastesper day with the total number of pilgrims per day in the 10 km2 area. A DON flux of606 kgha−1was calculated in the temple area where pilgrims conducted bathing and15

associated activities.High DON values were also observed in the downstream temple locations in the

plantation segment, particularly in the NEM. Because of the huge crowd many pil-grims go to downstream temple locations for bathing before visiting the shrine. Thewash off of nutrients from the human impacted locations upstream together with the20

anthropogenic activities in the location itself resulted in the second highest DON con-centration in location 6. Hence, the gathering of millions of people for the religiousevent and associated activities in the upstream catchment not only had an impact inlocations therein, but in the downstream segments as well. Therefore, the Sabarimalapilgrim event is considered as the major cause for the extremely high DON flux in the25

Pamba River.

16149

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

4.3 Impacts due to agriculture

DON accounts for a high percentage (99 %) of TDN concentrations in the agriculturaldominated segments in the Pamba catchment. When estimating segment wise, DONyield was high in the SMT and agricultural dominated segments (III to VI). A studyconducted in 7 contrasting agricultural land use types in England, Whales and Greece5

comprised an average DON of 57±8 % of the TDN pool in soil (Christou et al., 2005).Another agricultural dominated catchment in South-East Spain (Lorite-Herrera et al.,2009) accounted the DON percentage in the ground water (98.5 %) and river water(97.2 %) of the TDN pool and having concentrations ranging from 7 to 2442 µM and450 to 1414 µM, respectively. When compared with the Pamba catchment, the % com-10

position of DON in south-east Spain catchment was found similar (99 %), while theconcentrations (36 to 29 302 µM) were in a much lower range.

Generally, in the agricultural fields urea is often applied when the rainfall is antici-pated and about 3 to 5 % of the surface applied urea can be lost via runoff (Glibertet al., 2006). Following the general trend, fertilizer is applied before the onset of the15

SWM and NEM except for the paddy cultivated area, because this area in the Pambacatchment lies below sea level and usually gets flooded during heavy rains in the mon-soon months. To avoid this natural calamity, paddy cultivation normally begins duringthe end of NEM, i.e. during November. Fertilizer application starts by then and contin-ues in the PM.20

In the agricultural dominated Guadalquivir catchment, urea was applied at a rateof 350 kgha−1 (Lorite-Herrera et al., 2009). When compared to the Pamba catchment,the application rate was about 8-fold higher whereas, the DON yield in the Guadalquivircatchment was orders of magnitude lower (2 kgha−1). In the Guadalquivir catchmentsewage is not a significant source (Lorite-Herrera and Jiménez-Espinosa, 2008). Tak-25

ing into account the timing of fertilizer application, about 3 to 5 % of the surface appliedurea can be lost via runoff in the Pamba catchment. Considering the total amountof urea fertilizer applied in the agricultural fields (20 kgha−1 yr−1), probably less than

16150

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

1 kgha−1 yr−1 is lost via surface runoff, whereas the rest is taken up by the crops orretained in the soil in the Pamba catchment. Thus it is likely that urea application doesnot contribute significantly to the extremely high DON in the agriculture dominated seg-ments in the Pamba catchment.

Livestock excretions is another potential DON source in the agricultural dominated5

segments in the Pamba catchment. Livestock wastes contain extremely high DON con-centrations (Bristow et al., 1992). The N content in urine of grazing animals normallyranges from 8 to 15 gL−1. Urine increases the pH of the soil due to the hydrolysis ofurea (Haynes and Williams, 1993) and can considerably increase the DON and DOCtogether with other compounds in the soil solution (Shand and Coutts, 2006). The N10

concentrations in the soil exceed the ability of plants and soil organisims to utilise thatN for growth (Hoogendoorn et al., 2010). The stocking rates of animals per hectarehave a strong effect on the amount of urine deposited and subsequent DON losses(van Kessel et al., 2009). The soil surface nitrogen content was quantified for differentstates of India for the period 2000 to 2001 (Prasad et al., 2004). According to this study,15

livestock manure constituted a major percentage of total N inputs (44 %) in the agricul-tural regions. When compared to the other Indian states, N input from livestock excre-tions was high in Kerala thereby, comprising of 616.4 kgha−1. In the Pamba catchment,cattle and fowls are abundant (as per field observation) and is in accordance with thefindings of Prasad et al. (2004). Farmyard manure is commonly applied together with20

synthetic fertilizer in the catchment and amounted to 553 tyr−1of N in the agriculturalregion (Elizabeth David et al., 2013).

The total amount of N leached from livestock urine varies between 18 to 58 % ofthe N applied, depending on the soil texture (Clough et al., 1998). Taking into accountthe abovementioned range, in the Pamba catchment the DON leaching loss from live-25

stock excretions can range from 111 to 386 kgha−1 yr−1. For example, the DON lossfrom grazed grassland catchments in Northern Ireland ranged from 5 to 24 % of thetotal N and the soil therein is sandy-clay-loam (Watson et al., 2000). The soil in thePamba catchment is also sandy-clay-loam. Therefore, taking into account the soil tex-

16151

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

ture, a maximum of 24 % DON loss from livestock excretions is possible. The DON losscalculated via this source is 148 kgha−1 yr−1and makes up a significant portion of thetotal DON input in the Pamba River (Table 3).

Variations in the DON loading were observed in the agricultural dominated segments(III to VIII) in the Pamba River. DON loading in the agricultural dominated segments was5

lower during PM compared to the SWM and NEM (Fig. 4). The DON loads were highin segment II and III because of the maximum pilgrim activity during the PM. The highloads were not transported to downstream segments due to low flow conditions, conse-quently resulting in low DON load in the mid- and downstream agricultural dominatedsegments.10

The main drivers leading to DON loss in agricultural soils are precipitation and irriga-tion events and their frequencies (van Kessel et al., 2009). 70 % of the total precipitationis received during the SWM in the Pamba catchment. High DON loading at that timeis due to the washoff of agricultural soils as well as the N leaching via livestock excre-tions during high water flow conditions. In addition to this, heavy precipitation during the15

onset of SWM transports the DON that accumulated in the upstream temple segmentduring low flow conditions after the pilgrim event to the mid- and downstream segments.When compared to the SWM, the DON concentrations and load in the plantation andSMT segments during the NEM were higher which is mainly due to the excess nutrientloads from the pilgrim season as well as the N leaching loss via livestock excretions.20

When comparing the segment wise annual yield, DON decreased towards the rivermouth from segment VI, i.e. the yield was lower in the paddy cultivated segments VIIand VIII. Rice paddies favor denitrification in the wet climate (Borbor-Cordova et al.,2006). Urea applied in the field can be hydrolyzed to nitrate and under water loggedconditions due to microbial activity denitrification is enhanced, leading to the N loss25

thereby, resulting in low yields in the paddy cultivated segments in the Pamba catch-ment.

Most biological nitrogen fixation in terrestrial systems occurs in tropical regions (Gal-loway et al., 2008). River DON can also be derived from natural sources such as

16152

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

biological nitrogen fixation from agricultural crops (BNF-agr, Galloway et al., 2004).In the tropical Piracicaba basin in Brazil, the input from BNF-agr accounted for14.5 kgha−1 yr−1 (Filoso et al., 2003). For both the Oder and Vistula watersheds ofPoland, the input from BNF-agr amounted to 16 kgha−1 yr−1 each. In the Indian agricul-tural area, N fixation made up to 17.1 kgha−1 yr−1 (Prasad et al., 2004). It is therefore5

conceivable that the input from BNF-agr in the agricultural dominated Pamba water-shed adds significantly to the total DON load of the Pamba River. Atmospheric organicnitrogen (AON) deposition can also be a potential DON source in the Pamba catch-ment. Urea, amino acids, animal husbandary operations, agricultural and biomassburning accounts for the AON deposition, and comprise one third of the total N de-10

position (Neff et al., 2002). According to Galloway et al. (2008) the total N depositionrate for Kerala falls into the range 10–20 kgha−1 yr−1. Taking one third of this value ac-cording to Neff et al. (2002), results in about 5 kgha−1 yr−1of AON being deposited inthe Pamba catchment. In agricultural dominated catchments, where inorganic fertiliz-ers are applied on a regular basis, root exudation and turnover from different vegetation15

can also be a source of DON (Christou et al., 2006). Thus the impact of agriculture inthe Pamba catchment in terms of livestock farming is more significant compared to thefertilizer application in the fields thereby, contributing a major portion to the total DONinput in the Pamba River.

4.4 Impacts due to domestic waste disposal20

The delivery of waste water and organic fertilizer via surface runoff represent a poten-tial direct source of DON to streams in agricultural watersheds (Pellerin et al., 2006).A very common phenomenon in the Pamba catchment is that many houses on thebanks have their sewage pipes turned to the river thereby, facilitating flow of domes-tic sewage into it. Domestic sewage includes waste water from bathing, washing, food25

preparation and from dish washing along with faeces and urine. On average, organicN in domestic sewage per person per day is 3.4 g and ammonia and urea N comprises

16153

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

5.3 g, respectively (Painter and Viney, 1959). In waste water approximately 80 % of thetotal N is urea (Hanson and Lee, 1971).

Population density increases from the upstream to the downstream segments in thePamba catchment. Domestic sewage inputs in these segments were high because,more people live near the river and the population density on a micro-scale is more than5

in the other parts of the catchment. Annual domestic sewage N inputs in the Pambacatchment were calculated by multiplying the population density per square kilometerwith the amount of domestic sewage N produced per person per day. Approximately,26 kgha−1 yr−1of N was introduced into the river via domestic sewage inputs and wasan order of magnitude higher than the other human impacted world rivers (Fig. 5b).10

Domestic sewage inputs despite being high, the values does not balance the highDON yields in the mid-and downstream segments of the Pamba River. DON yields inthese segments were much lower than for the temple segments. This indicates that thedomestic sewage inputs from the settlement region were not high as the inputs fromthe pilgrims (Fig. 5c). Therefore, the yields in these segments are the cumulative effect15

of the whole upstream river catchment. DON uptake by phytoplankton and bacteriamight be the reasons for the much lower DON yield in the paddy segments in contrastto high domestic sewage inputs due to high population density therein.

Water resource management is a major challenge in India and the country has highincident and adjusted human water security threat. The semi-arid climate and high sea-20

sonal water availability coupled with pollution and water consumption by the large andgrowing population resulted in high incident threat whereas, the investment in tech-nology and infrastructure to improve human access to water resources is low therebyleading to adjusted human water security threat in India. The investment level in waterresource infrastructure varies among parts of the country, but is rarely high, leaving25

most of the country with high adjusted threat relative to other regions in the world(Vörösmarty et al., 2010) The lack of infrastructure, in particular for the sewage col-lection and treatment, is most probably the reason that the sewage N in the Pamba

16154

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

catchment was an order of magnitude higher than in other human impacted catch-ments (Fig. 5).

4.5 Dissolved organic nitrogen budgets in the global context

The Pamba River transported 128 061 t of DON to the receiving Vembanad estuary,annually. The Pamba DON yield amounted to 745 kgha−1 yr−1and was two orders of5

magnitude higher compared to results from the NEWS (Global Nutrient export fromWatersheds, Fig. 5) DON model (Harrison et al., 2005). The DON yield increasedlinearly with population density (r2 = 0.786, p < 0.001, Fig. 5a) in the Pamba catch-ment. Without the Pamba data, the DON yield was only weakly correlated with pop-ulation density in other world river catchments (r2 = 0.0514, p < 0.001). Sewage N10

input in the Pamba catchment was almost an order of magnitude higher than calcu-lated by the NEWS DON model (Fig. 5b) and exhibited a significant positive correlationwith the population density (r2 = 0.722, p < 0.001) and with the DON yield (r2 = 0.840,p < 0.001). When excluding the Pamba data, sewage N inputs from other catchmentswere weakly correlated with both the population density (r2 = 0.366, p < 0.001) and15

DON yield (r2 = 0.0376, p < 0.001). The NEWS DON model suggests that the diffuseanthropogenic source, i.e. from agricultural fields accounts for 11 %, while the pointsource, i.e. from sewage accounts for 3 % of DON, respectively (Fig. 6). For the PambaRiver catchment it is concluded that the high population density and associated humanactivities together with the lack of infrastructure for domestic waste disposal in the set-20

tlement region as well as from the unique Sabarimala pilgrimage are major factors forthe extremely high DON yield.

We quantified the inputs of DON in the Pamba catchment from the pilgrim activ-ity, agriculture and sewage effluents (Table 3). As discussed before a DON input of606 kgha−1 was calculated from the 10 km2 temple area and when normalized to the25

whole catchment, the input from the pilgrim activity amounted to 271 kgha−1 yr−1. Itcomprised of 53 % of the total DON inputs and 36 % of the total DON yield for thePamba catchment. Domestic sewage N inputs and urea fertilizer amounted to 26 and

16155

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

< 1 kgha−1 yr−1 resulting in 3 and < 1 % of the total catchment yield, respectively.Therefore, inputs from all the above mentioned absolute sources made up to 40 %of the total DON yield in the Pamba catchment (Table 3).

The major potential DON input quantified for the Pamba River is the N leaching lossvia livestock excretion. 148 kgha−1 yr−1 stem from this source and made up the second5

largest DON input (29 %) after the pilgrim activity (53 %). It amounts to 20 % of the totalDON yield. Atmospheric organic N deposition in the Pamba catchment amounted to5 kgha−1 yr−1 and made up 1 % of the total catchment yield in the Pamba River. Theother sources such as precipitation, throughfall, soil solution and BNF-agr contribute2 % each to the total Pamba River DON yield. The abovementioned DON sources to-10

gether made up 29 % of the total DON yield (Table 3) and contribute to a significantamount to the extremely high DON flux in the Pamba catchment. The total DON in-puts from all sources amounted to 514 kgha−1 yr−1 (range 477 to 752 kgha−1 yr−1) andaccount for 69 % (69 to 101 %) of the total measured DON in the Pamba catchment.The unaccounted 31 % are probably the result of (i) uncertainties in estimated terms15

for which no field data are available and (ii) missing data from regions with comparableenvironmental settings even not allowing for an estimate. For example, this is the casefor BNF in the forest region and root exudation from inorganic fertilizer application inthe agricultural land use area.

5 Summary and conclusions20

Anthropogenic factors like the high population density, sewage disposal and other activ-ities such as agriculture and livestock farming are responsible for the global maximumDON concentrations and fluxes measured in the Pamba River. In contrast to manyother rivers, DON not only is the dominant form of N, but accounts for 99 % of theTDN pool in the Pamba River. A specific cause-effect relationship was identified in the25

Pamba catchment by quantifying the DON fluxes from respective land use segments.The unique Sabarimala pilgrimage was the major reason of the extreme DON load and

16156

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

yield in the Pamba River and is due to the lack of infrastructure for sanitation and wastedisposal. In the Pamba River, the high nutrient load during the pilgrim event can be re-duced to a major extent if there were sufficient toilets to suffice the need of the hugenumber of pilgrims visiting the temple.

A high threat to human water security and river biodiversity are a major challenge5

in the densely populated regions of the poor world countries, particularly in South andSoutheast Asia and parts of Africa. This is because of the lack of investments in tech-nology for clean drinking water and sanitation (Vörösmarty et al., 2010). From our studyit is evident that the lack of technological investments not only affects the drinking waterand river biodiversity, but also has a huge impact on the riverine nutrient concentra-10

tions. Therefore, the DON derived from sewage can be a major problem to the coastalwaters rather than DIN in densely populated river catchments. Our results suggest thatinstallation of adequate sanitation measures for waste disposal and treatment togetherwith improved agricultural and animal management practices, could strongly reducethe anthropogenic N input into the Pamba River. Our study also underscores the need15

for more regional scale quantitative studies that assess the effects of anthropogenicactivities on riverine DON fluxes in tropical regions in order to improve water qualitymanagement and nitrogen budget calculations.

Acknowledgements. We thank Srikumar Chattopadhyay and Mahamaya Chattopadhyay, Sci-entists, Centre for Earth Science Studies (CESS), Kerala, India for the enormous help during20

the course of the study. We thank Aneesh R, Arun R, Lijith P, Resmi Sajan, Vandana M of CESSand Tz-Ching Yeh for the help during field work and map preparations. Thanks to Matthias Bir-kicht for the assistance in the laboratory at the ZMT. Special thanks to Ulrike Machulik foranalyzing the samples in the Institute of Biogeochemistry and Marine Chemistry, University ofHamburg. Financial support by the ZMT and the German Academic Exchange Service (DAAD)25

is gratefully acknowledged.

16157

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

References

Bhardwaj, R.: Water quality monitoring in India-achievements and constraints, IWG-Env, Inter-national Work Session on Water Statistics, Vienna, 1–12, 2005.

Borbor-Cordova, M. J., Boyer, E. W., McDowell, W. H., and Hall, C. A.: Nitrogen and phospho-rus budgets for a tropical watershed impacted by agricultural land use: Guayas, Ecuador,5

Biogeochemistry, 79, 135–161, 2006.Bouwman, A. F., Van Drecht, G., Knoop, J. M., Beusen, A. H. W., and Meinardi, C. R.: Exploring

changes in river nitrogen export to the world’s oceans, Global Biogeochem. Cy., 19, GB1002,doi:10.1029/2004GB002314, 2005.

Boyer, E. W., Howarth, R. W., Galloway, J. N., Dentener, F. J., Green, P. A., and Vörös-10

marty, C. J.: Riverine nitrogen export from the continents to the coasts, Global Bio-geochem.Cy., 20, GB1S91, doi:10.1029/2005GB002537, 2006.

Bristow, A. W., Whitehead, D. C., and Cockburn, J. E.: Nitrogenous constituents in the urine ofcattle, sheep and goats, J. Sci. Food Agr., 59, 387–394, doi:10.1002/jsfa.2740590316, 1992.

BusinessLine: Medical and sanitary facilities at Sabaraimala “woefully inadequate”, The Hindu15

group of Publications, 2009.Campbell, J., Hornbeck, J., McDowell, W., Buso, D., Shanley, J., and Likens, G.: Dissolved

organic nitrogen budgets for upland, forested ecosystems in New England, Biogeochemistry,49, 123–142, doi:10.1023/a:1006383731753, 2000.

Census of India: Office of Registrar General & Census Commissioner, available at:20

www.censusindia.gov.in, 2011.Chattopadhyay, S. and Chattopadhyay, M.: Impact of terrain characteristics on catchment yield

– a case of Kabbini River Basin, Kerala, in: Perspectives in Resource Management in De-veloping Countries – Land Appraisal and Development, edited by: Thakur, B., Concept’sInternational Series in Geography, Concept Publishing Co., New Delhi, 4, 145–157, 2009.25

Christou, M., Avramides, E. J., Roberts, J. P., and Jones, D. L.: Dissolved organicnitrogen in contrasting agricultural ecosystems, Soil Biol. Biochem., 37, 1560–1563,doi:10.1016/j.soilbio.2005.01.025, 2005.

Christou, M., Avramides, E. J., and Jones, D. L.: Dissolved organic nitrogen dy-namics in a Mediterranean vineyard soil, Soil Biol. Biochem., 38, 2265–2277,30

doi:10.1016/j.soilbio.2006.01.025, 2006.

16158

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

Clough, T. J., Ledgard, S. F., Sprosen, M. S., and Kear, M. J.: Fate of 15N labelled urine on foursoil types, Plant Soil, 199, 195–203, doi:10.1023/a:1004361009708, 1998.

Cole, J., Peierls, B., Caraco, N., and Pace, M.: Nitrogen loading of rivers as a human-drivenprocess, in: Humans as Components of Ecosystems, edited by: McDonnell, M. and Pick-ett, S. A., Springer, New York, 141–157, 1993.5

Elizabeth David, S., Chattopadhyay, M., Chattopadhyay, S., and Jennerjahn, T. C.: Sci. TotalEnviron., in review, 2013.

Fang, Y. T., Gundersen, P., Mo, J. M., and Zhu, W. X.: Input and output of dissolved organicand inorganic nitrogen in subtropical forests of South China under high air pollution, Biogeo-sciences, 5, 339–352, doi:10.5194/bg-5-339-2008, 2008.10

Filoso, S., Martinelli, L., Williams, M., Lara, L., Krusche, A., Victoria Ballester, M., Victoria, R.,and de Camargo, P.: Land use and nitrogen export in the Piracicaba River basin, SoutheastBrazil, Biogeochemistry, 65, 275–294, doi:10.1023/a:1026259929269, 2003.

Filoso, S., Martinelli, L., Howarth, R., Boyer, E., and Dentener, F.: Human activities changingthe nitrogen cycle in Brazil, in: Nitrogen Cycling in the Americas: Natural and Anthropogenic15

Influences and Controls, edited by: Martinelli, L. and Howarth, R., Springer, the Netherlands,61–89, 2006.

Galloway, J. N., Dentener, F. J., Capone, D. G., Boyer, E. W., Howarth, R. W., Seitzinger, S. P.,Asner, G. P., Cleveland, C. C., Green, P. A., Holland, E. A., Karl, D. M., Michaels, A. F.,Porter, J. H., Townsend, A. R., and Vöosmarty, C. J.: Nitrogen cycles: past, present, and20

future, Biogeochemistry, 70, 153–226, doi:10.1007/s10533-004-0370-0, 2004.Galloway, J. N., Townsend, A. R., Erisman, J. W., Bekunda, M., Cai, Z., Freney, J. R.,

Martinelli, L. A., Seitzinger, S. P., and Sutton, M. A.: Transformation of the nitro-gen cycle: recent trends, questions, and potential solutions, Science, 320, 889–892,doi:10.1126/science.1136674, 2008.25

Glibert, P., Harrison, J., Heil, C., and Seitzinger, S.: Escalating worldwide use of urea –a global change contributing to coastal eutrophication, Biogeochemistry, 77, 441–463,doi:10.1007/s10533-005-3070-5, 2006.

Grasshoff, K., Kremling, K., and Ehrhardt, M. (Eds.): Methods of Seawater Analysis, 3rd edn.,Wiley-VCH, Weinheim, 1999.30

Hanson, A. M. and Lee, G. F.: Forms of organic nitrogen in domestic wastewater, J. WaterPollut. Control Fed., 43, 2271–2279, http://www.jstor.org/stable/25037233, 1971.

16159

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

Harrison, J. A., Caraco, N., and Seitzinger, S. P.: Global patterns and sources of dissolvedorganic matter export to the coastal zone: Results from a spatially explicit, global model,Global Biogeochem. Cy., 19, doi:10.1029/2005gb002480, 2005.

Haynes, R. J. and Williams, P. H.: Nutrient cycling and soil fertility in the grazed pasture ecosys-tem, in: Advances in Agronomy, edited by: Donald, L. S., Academic Press, 119–199, 1993.5

Hoogendoorn, C. J., Betteridge, K., Costall, D. A., and Ledgard, S. F.: Nitrogen concentrationin the urine of cattle, sheep and deer grazing a common ryegrass/cocksfoot/white cloverpasture, New Zeal. J. Agr. Res., 53, 235–243, doi:10.1080/00288233.2010.499899, 2010.

Jordan, T. E. and Weller, D. E.: Human contributions to terrestrial nitrogen flux, BioScience, 46,655–664, doi:10.2307/1312895, 1996.10

Lorite-Herrera, M. and Jiménez-Espinosa, R.: Impact of agricultural activity and geologic con-trols on groundwater quality of the alluvial aquifer of the Guadalquivir River (province ofJaén, Spain): a case study, Environ. Geol., 54, 1391–1402, doi:10.1007/s00254-007-0920-6, 2008.

Lorite-Herrera, M., Hiscock, K., and Jiménez-Espinosa, R.: Distribution of dissolved inorganic15

and organic nitrogen in river water and groundwater in an agriculturally-dominated catch-ment, south-east Spain, Water Air Soil Pollut., 198, 335–346, doi:10.1007/s11270-008-9849-y, 2009.

Lovett, G. M., Weathers, K. C., and Sobczak, W. V.: Nitrogen saturation and retention in forestedwatersheds of the Catskill Mountains, New York, Ecol. Appl., 10, 73–84, 2000.20

Martinelli, L., Coletta, L., Ravagnani, E., Camargo, P., Ometto, J., Filoso, S., and Victoria, R.:Dissolved nitrogen in rivers: comparing pristine and impacted regions of Brazil, Braz. J. Biol.,70, 709–722, 2010.

Neff, J. C., Holland, E. A., Dentener, F. J., McDowell, W. H., and Russell, K. M.: The origin,composition and rates of organic nitrogen deposition: a missing piece of the nitrogen cycle?,25

Biogeochemistry, 57, 99–136, 2002.Painter, H. A. and Viney, M.: Composition of a domestic sewage, J. Biochem. Microbiol., 1,

143–162, doi:10.1002/jbmte.390010203, 1959.Pellerin, B., Kaushal, S., and McDowell, W.: Does anthropogenic nitrogen enrichment increase

organic nitrogen concentrations in runoff from forested and human-dominated watersheds?,30

Ecosystems, 9, 852–864, doi:10.1007/s10021-006-0076-3, 2006.Perakis, S. S. and Hedin, L. O.: Nitrogen loss from unpolluted South American forests mainly

via dissolved organic compounds, Nature, 415, 416–419, 2002.

16160

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

Prasad, V. K., Badarinath, K. V. S., Yonemura, S., and Tsuruta, H.: Regional inventory of soilsurface nitrogen balances in Indian agriculture (2000–2001), J. Environ. Manage., 73, 209–218, doi:10.1016/j.jenvman.2004.06.013, 2004.

Rabalais, N. N.: Nitrogen in aquatic ecosystems, AMBIO, 31, 102–112, doi:10.1579/0044-7447-31.2.102, 2002.5

Schwendenmann, L. and Veldkamp, E.: The role of dissolved organic carbon, dissolved organicnitrogen, and dissolved inorganic nitrogen in a tropical wet forest ecosystem, Ecosystems,8, 339–351, doi:10.1007/s10021-003-0088-1, 2005.

Seitzinger, S. and Sanders, R.: Contribution of dissolved organic nitrogen from rivers to estuar-ine eutrophication, Mar. Ecol.-Prog. Ser., 159, 1–12, 1997.10

Seitzinger, S. P., Sanders, R., and Styles, R.: Bioavailability of DON from natural and anthro-pogenic sources to estuarine plankton, Limnol. Oceanogr., 47, 353–366, 2002.

Seitzinger, S. P., Hartnett, H., Lauck, R., Mazurek, M., Minegishi, T., Spyres, G., and Styles, R.:Molecular-level chemical characterization and bioavailability of dissolved organic matter instream water using electrospray-ionization mass spectrometry, Limnol. Oceanogr., 50, 1–12,15

2005.Shand, C. A. and Coutts, G.: The effects of sheep faeces on soil solution composition, Plant

Soil, 285, 135–148, doi:10.1007/s11104-006-0065-5, 2006.Turner, R. E., Rabalais, N. N., Justic, D., and Dortch, Q.: Global patterns of dissolved N, P and

Si in large rivers, Biogeochemistry, 64, 297–317, doi:10.1023/a:1024960007569, 2003.20

Tyagi, V. K., Bhatia, A., Gaur, R. Z., Khan, A. A., Ali, M., Khursheed, A., Kazmi, A. A.,and Lo, S.-L.: Impairment in water quality of Ganges River and consequential healthrisks on account of mass ritualistic bathing, Desal. Water Treatm., 51, 2121–2129,doi:10.1080/19443994.2013.734677, 2012.

van Kessel, C., Clough, T., and van Groenigen, J. W.: Dissolved organic nitrogen: an over-25

looked pathway of nitrogen loss from agricultural systems?, J. Environ. Qual., 38, 393–401,doi:10.2134/jeq2008.0277, 2009.

Vitousek, P. M., Aber, J. D., Howarth, R. W., Likens, G. E., Matson, P. A., Schindler, D. W.,Schlesinger, W. H., and Tilman, D. G.: Human alterations of the global nitrogen cycle: sourcesand consequences, Ecol. Appl., 7, 737–750, 1997.30

Vorosmarty, C. J., McIntyre, P. B., Gessner, M. O., Dudgeon, D., Prusevich, A., Green, P., Glid-den, S., Bunn, S. E., Sullivan, C. A., Liermann, C. R., and Davies, P. M.: Global threats tohuman water security and river biodiversity, Nature, 467, 555–561, 2010.

16161

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

Watson, C. J., Jordan, C., Lennox, S. D., Smith, R. V., and Steen, R. W. J.: Organic nitrogenin drainage water from grassland in Northern Ireland, J. Environ. Qual., 29, 1233–1238,doi:10.2134/jeq2000.00472425002900040029x, 2000.

Westerhoff, P. and Mash, H.: Dissolved organic nitrogen in drinking water supplies: a review,Aqua, 51, 415–448, 2002.5

WHO/UNICEF: Progress on Sanitation and Drinking-Water: 2010 Update, Joint Monitoring Pro-gramme for Water Supply and Sanitation (World Health Organisation/UNICEF, 2010), 2010.

16162

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

Table 1. Land use categories in the Pamba catchment in 2010.

Landuse Area in km2 % of total area

Forest 977.6 43.7Settlement with mixed tree crops (SMT) 318.9 14.3Rubber 247.6 11.1Paddy 210.4 9.4Water body 139.4 6.2Fallow land 132.9 5.9Settlements 88.2 3.9Tea 33.7 1.5Marshy land 22.5 1.0Open scrub 20.5 0.9Reservoir 20.0 0.9Eucalyptus 16.4 0.7Cardamom 7.7 0.3

Total catchment area 2235

16163

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

Table 2. Segment wise annual DON load and yield for the Pamba River.

Area Segment Discharge Population density DON load DON yieldSegment (km2) Dominant land use area (%) (km3 yr−1) (inhabitantskm−2) (t yr−1) (kgha−1 yr−1)

I 249.6 Forest 11 0.4 < 1 2244 90II 219.7 Temple, Forest 10 0.4 < 1 43 398 1976III 359.5 Forest, Plantation 16 0.6 78 66 982 1863IV 621.1 Forest, Rubber, Settlement 28 1.1 148 84 929 1367V 110.1 Rubber, Smt 5 0.2 583 20 017 1818VI 159.8 Smt 7 0.3 894 21 925 1372VII 175.3 Smt, Paddy 8 0.3 1313 13 421 765VIII 339.8 Paddy, Smt (Wetland) 15 0.6 1034 22 546 663

16164

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

Table 3. Dissolved organic nitrogen inputs into the Pamba River.

DON % of the % of theSources (kgha−1 yr−1) total input total DON yield References

Pilgrims 271 53 36 This studyDomestic sewage 26 5 3 This studyUrea < 1 < 1 < 1 This studyLivestock excretion 148 29 20 Prasad et al. (2004)Precipitation 18 3 2 Fang et al. (2008)Throughfall (average) 17 3 2 Fang et al. (2008)BNF-agr 17 3 2 Prasad et al. (2004)Soil solution (average) 12 2 2 Fang et al. (2008)Atmospheric organic N 5 1 1 Neff et al. (2002)(AON) deposition

Total inputs (kgha−1 yr−1) 514 (477 to 752 kgha−1 yr−1)

16165

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

Fig. 1. Hydrological cycle, fertilization and pilgrim events in the Pamba basin. Average pre-cipitation (grey bars) and discharge (filled circles, solid lines) pattern in Alappuzha andPathanamthitta districts during the period 2000–2005. Precipitation during 2010 campain (filledtriangles).

16166

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

Fig. 2. Land use map of the Pamba River including sampling locations and segments.

16167

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

Fig. 3. Dissolved organic nitrogen concentrations during the premonsoon, southwest monsoonand northeast monsoon seasons in the Pamba catchment.

16168

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

Fig. 4. Dissolved organic nitrogen loads during the premonsoon, southwest monsoon andnortheast monsoon seasons in the Pamba catchment.

16169

Discussion

Paper

|D

iscussionP

aper|

Discussion

Paper

|D

iscussionP

aper|

Fig. 5. Relationship between (a) population density and DON yield; (b) population density andsewage input; (c) sewage input and DON yield. Data for the comparison are taken from theNEWS (Global Nutrient export from Watersheds) DON model by Harrison et al. (2005). Thesolid line denotes the linear regression line for all data points, the dashed line denotes it for alldata points excluding the Pamba River.

16170

Discussion

Pa

per|

Discussion

Pa

per|

Discussion

Paper

|D

iscussionP

aper|

Fig. 6. Comparison of DON inputs in the Pamba Riverwith other world rivers. Input from sewage(filled bars) and manure (dotted bars). Data for other rivers are from Harrison et al. (2005).

16171